This article will be a summary of the Standings Gain Points (SGP) approach to Roto fantasy baseball and I will use the TGFBI (The Great Fantasy Baseball Invitational) as the draft I used it for to show you how.
There are many ways to go about completing a fantasy draft but most of the systems out there are designed around taking player projections and assigning values to them. They all then rank the players, in different ways, to give you an auction value which you can use in a fantasy auction or as an order for a snake draft.
SGP is built off the idea that you know how the league has performed historically and you take your projections for the new season and award points accordingly.
So first you need to have some projections for the upcoming season. The are multiple projection systems out there, with there being many free on FanGraphs as well as many others which are available for small sums of money (Projecting X 2.0, Razzball) or part of other subscriptions (Baseball Prospectus).
I merge these together to produce one projection which has average totals and the variance of the performance between the system. With this data I can start to use the SGP model to value the players. If you want to make any manual changes to the projections, you should perform them here not later on.
The end goal of rotisserie (roto) fantasy baseball is to accumulate the most points in the standings. The Standings Gain Points approach to valuing players is to convert a player’s statistics into the number of rotisserie points those statistics are worth.
In order to do this, you must have some historical data for the league, or similar leagues. In 2019 the TGBFI has changed to the NFBC Main Event format, which is a 15-team standard 5×5 roto league with 2 catchers.
Now you might think that since this is a new format, how could I calculate the historic performance? Well, thanks to this being on NFBC, the data on previous tournaments is available so I can take the data from the 2018 seasons and use that as an approximation for 2019. These leagues require some of the highest entry fees and winnings out there so they would be a good comparison to a league built of fantasy experts and Tom Pringle.
Armed with this data I can get to see what was average for each position in a 15 team league with the same rules in 2018.
Now we have this we can calculate the increase in each category required to jump up 1 place and score one point. You can go about this in several different methods, but they all boil down to looking at the difference between the middle teams and using them as your valuation of 1 point.
From the data above I surmised the below as the amounts required to move up one position.
With these values I can determine the value of each player in 2019 in a 15 team 5×5 league with 2 catchers. It is important to know that when you move on from this step your rankings are now unique to the league you are doing it for and don’t translate directly for different leagues.
For the counting stats (R, RBI etc) you calculate the value for each player by just dividing their projected value by 1-point valuation. E.g. a player with 33 projected home runs has a valuation of 6 (33/5.5 = 6) in the home run category.
For the rate stats it is a bit more complicated. You will need to account for how much each player will contribute (ability and amount) to the overall and how much the average player would contribute. The change to the overall base rate of this player is then divided by the 1-point value to give the value for that category.
If the world was nice and simple, I have could just ended here, ranked the players accordingly and used them for my snake draft. But we must consider something else, the level of replacement players.
In TGFBI, a roster consists of 23 starters and 7 reserves for a team of 30 players. The starters are spilt with 9 pitchers and 14 hitters, including 2 catchers.
To understand the value of a player we must consider the value of the player we could replace them with via free agency. These would be the players with the highest value that we would expect to be outside of the main rosters. We call this replacement level and we remove the value of these players from all players of that position to get the value above replacement (VAR).
For example, there are positions for 30 catchers across the 15 teams meaning the that the value of the 31st catcher is the replacement level for catchers. This should be calculated for all positions, and the value would be different from the replacement value of pitchers or outfielders.
Armed with the replacement level data we can calculate the VAR of each player so if we compare our 4 players from before their value changes differently.
Sale and Scherzer’s values are now much closer to Trout and Betts because the replacement level in much lower for pitchers compared to outfielders. This also significantly brings up the value of catchers: Realmuto and Sanchez move up 70 places into the top 40, as the replacement level is much lower than all other hitting positions.
I now have my rankings for the TGFBI draft but there is one further piece of information we require before being ready for the draft. That is the ADP (Average Draft Position) of the players in similar drafts that have happened so far. This can be obtained from multiple places but thankfully NFBC has ones for drafts like my upcoming one to choose from.
We want this information to highlight differences between our valuation and the public’s valuation for one reason. Can we pass on players and still expect to get them in a future round? This generally only applies to players after the 4th or 5th round as that is where differences in opinion might get to the size of one round.
The reasoning behind this is thus: imagine you have a player who you rank 80 but the ADP rankings are saying the player is going on average around 150. If you take the player at 80 when you next draft, you are missing out on potential value by drafting the player at your projected cost when he could be had at a discount of multiple rounds.
If you have multiple players you feel are undervalued then you should just work through as many of them as you can because you won’t get them all.
With this information, I was ready for the TGFBI draft. I have uploaded my overview which you can view here. This was accurate as of the start of the draft on the 24th Feb.
NFBC has one further detail with the draft which is different to some leagues: it has the Kentucky Derby System (KDS) for draft preferences. That means that you aren’t randomly assigned a place in the draft. You choose an order and when you get assigned a spot it chooses the highest in your preference still left.
My plan for the draft was to try and take a five-tool hitter first. So, I left 1-4 at the start of my preference and put 14 & 15 next hoping to get a top power hitter and ace if placed here.
I ended up with the 4th spot so I roughly got what I wanted. I will go into details for the draft in another post shortly.
If you want more details on SGP and building your own system I highly recommend “The Process” by Jeff Zimmerman and Tanner Bell. It goes into the method in much greater detail as well as hints for other projection methods and in-season tips.